Summer School in Statistics 2010
Time Series Modeling with ARIMA and RegARIMA Models
A Summer School co-arranged by Örebro University and Statistics Sweden, Örebro, August 9-11, 2010
A time series is a sequence of data points, each of which represents the value of the same variable at different times, usually measured at equidistant time intervals. Examples are found in economics (price indices, unemployment measurements, production indices), finance (stock exchange rate, stock market index), meteorology (temperature and rainfall records), and in many other areas.
Time series are often thoroughly scrutinized, as they can provide important information on the development of e.g. the economy, the stock market or the climate. For this purpose, statistical models for analysis of time series data are important tools and several modeling approaches are available.
For the purpose of forecasting or of estimating seasonal and other calendar effects, regression models whose error components are modeled by autoregressive integrated moving average models have been found to be very effective.
For application of these regARIMA models, as they are called, there are several freely downloadable software packages that are highly developed to simplify the task of modeling successfully. For example, these packages have automatic methods for finding an initial model and for finding and accounting for outlying data points that could compromise the goal of modeling if ignored.
For any model being considered, they also have extensive model quality diagnostics, including graphical diagnostics.
Contents
Speakers
Details and registration
